Today we are moving Science Clouds to its own web pages. In addition to enabling quite a few exploratory projects, the Science Clouds to date served as a bit of a “cloud clinic” where various folks interested in using cloud computing for a scientific project would contact us and get advice and help on how to get started using Infrastructure-as-a-Service (IaaS) clouds.
Over time, these efforts resulted in shared images, papers, cloud evaluation projects, and other endeavors of which the following had the most impact:
The STAR “last minute” experiment: one of our most fascinating experiences last year has been helping the scientists of the STAR nuclear physics experiment meet a conference deadline by conducting a significant last-minute run on Amazon’s EC2. The run took place over 300+ virtual nodes, deployed as virtual clusters with the help of the Nimbus Context Broker, and ran over roughly 10 days consuming more than 36,000 hours of compute time. More details and perspectives are available in articles in ISGTW, HPCwire, and Newsweek.
The ALICE project with CernVM: the challenge was to elastically extend the globally distributed set of resources available to the ALICE HEP experiment — one of the four Large Hydron Collider experiments at CERN — in such a way that users don’t even notice whether their code runs in cloud provided resources or not. This was achieved by sensing resource demand and dynamically deploying virtual machine images developed by the CernVM project. More details available in HPCwire.
Sky Computing: can you create an environment deployed over a federation of cloud resources that is configured and protected in the same way as a local cluster? We defined and evaluated it in a paper that appeared in the September/October 2009 IEEE Internet Computing issue.
We appreciate all efforts and ideas that we have seen in the last year. With this new site we’d like to broaden the potential for discussions and share not just resources and images, but also papers, thoughts, and ideas on all sorts of topics related to how cloud computing can help science. If you have ideas, experiences or thoughts, let us know.